Law knowledge auxiliary learning method and system based on big data
By building generators and discriminators, and improving clustering algorithms by defining multiple fusion distances and designing clustering center update mechanisms, the problems of low data processing efficiency and poor clustering effect in traditional legal knowledge-assisted learning methods are solved, achieving higher quality data generation and more accurate clustering results.
Patent Information
- Application Number
- CN202411957644.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-30
- Publication Date
- 2025-04-25
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Traditional legal knowledge-assisted learning methods have problems such as low data processing efficiency, low data quality, and difficulty in effectively generating high-quality learning data. They also have poor clustering effects, difficulty in accurately reflecting the complex relationships between data vectors, inflexible clustering center update mechanism, and inappropriate processing of abnormal data.
By building generators and discriminators, the loss function is designed to improve the quality of generated data, define multivariate fusion distances, calculate matching degrees, design clustering and improve the methods of clustering by setting exception detection distances and clearing exception vectors.
It improves the diversity and representativeness of data, alleviates the problem of data scarcity, enhances the ability of the clustering algorithm to identify complex relationships between data vectors, avoids interference with the clustering results by abnormal data, improves the accuracy and stability of clustering, and enhances the adaptability of the clustering algorithm.
Smart Images
Figure CN119887467A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of intelligent assisted learning technology, and specifically to a legal knowledge assisted learning method and system based on big data. Background Art
[0002] The legal knowledge assisted learning method and system based on big data is an innovative tool that uses big data technology to assist in legal knowledge learning. It can identify the learning situation of learners, intelligently push relevant learning resources and monitor the learning status, effectively improving the efficiency and depth of legal learning.
[0003] Traditional legal knowledge-assisted learning methods have problems such as low data processing efficiency, low data quality, and difficulty in effectively generating high-quality learning data; traditional legal knowledge-assisted learning clustering methods have problems such as poor clustering effect, difficulty in accurately reflecting the complex relationship between data vectors, inflexible cluster center update mechanism, and inappropriate processing of abnormal data. Summary of the invention
[0004] In view of the above situation, in order to overcome the defects of the prior art, the present invention provides a method and system for assisting learning of legal knowledge based on big data. In view of the problems of low data processing efficiency, low data quality and difficulty in effectively generating high-quality learning data in traditional legal knowledge assisting learning methods, this solution improves the quality of generated data by constructing a generator and a discriminator and designing a loss function, thereby improving the diversity and representativeness of data and alleviating the problem of data scarcity. In view of the problems of poor clustering effect, difficulty in accurately reflecting the complex relationship between data vectors, inflexible cluster center update mechanism and inappropriate abnormal data processing in traditional legal knowledge assisting learning clustering methods, this solution improves clustering by defining multivariate fusion distance, calculating matching degree, designing cluster center update mechanism, setting abnormal detection distance and clearing abnormal vectors. The clustering algorithm is improved in its ability to identify complex relationships between data vectors, can adapt to changes in data vectors more quickly, avoids the interference of abnormal data on clustering results, improves the accuracy and stability of clustering, and enhances the adaptability of clustering algorithms.
[0005] The technical solution adopted by the present invention is as follows: a legal knowledge auxiliary learning method based on big data, the method comprising the following steps:
[0006] Step S1: data collection;
[0007] Step S2: data processing;
[0008] Step S3: constructing a legal knowledge-assisted learning clustering model;
[0009] Step S4: Legal knowledge assisted learning.
[0010] Furthermore, in step S1, the data collection is to collect students' grade, learning behavior data, legal ability data, legal knowledge learning status data and legal knowledge resources; the learning behavior data includes learning time, learning frequency and number of discussions participated in; the legal ability data includes homework completion status and test scores; the legal knowledge learning status data refers to the students' learning progress of legal knowledge points; the legal knowledge resources are textbooks, courses, papers, cases, legal provisions and legal classics corresponding to each legal knowledge point.
[0011] Furthermore, in step S2, the data processing specifically includes the following steps:
[0012] Step S21: constructing a data vector, constructing the collected data into a data vector in the format of: the student's grade, study time, study frequency, number of discussions, homework completion, test scores, and learning progress of legal knowledge points;
[0013] Step S22: construct a generator, the generator includes an input layer, a hidden layer and an output layer; the input layer receives a noise vector; the hidden layer performs an upsampling operation; the output layer uses a hyperbolic tangent function to map the output;
[0014] Step S23: construct a discriminator, which includes a convolution layer, a fully connected layer and an output layer; the convolution layer performs a one-dimensional convolution operation and then maps the output by a linear rectification function; the fully connected layer maps the extracted features to the output space to reduce the dimension of the input data; the output layer uses a sigmoid function to output a value between 0 and 1, indicating the probability that the input data is real data;
[0015] Step S24: Design the generator loss function, which is expressed as follows:
[0016] ;
[0017] in, represents the loss value of the generator, represents the simulation data generated by the generator, represents a generator, Represents the generator based on noisy data and the conditional vector The generated simulation data represents the distribution of simulation data generated by the generator, represents taking the logarithm, represents the discriminator, It means taking the expected value. and represents the generation loss weight, l represents the layer index of the discriminator, Indicates that the discriminator uses simulated data at layer l The extracted features, It means that the discriminator is trained on the real data at layer l The extracted features, It means taking the square of L2 norm, represents the distribution of real data, Indicates that the discriminator is inputting real data And the conditional vector The identification result at Indicates that the discriminator is inputting simulated data And the conditional vector The identification result at that time;
[0018] Step S25: Design the discriminator loss function, expressed as follows:
[0019] ;
[0020] in, represents the loss value of the discriminator, represents the identification loss weight, d represents the dimension index of the data vector, and D represents the maximum dimension of the data vector. Represents the value of the real data in the dth dimension, Represents the value of the simulation data in the dth dimension;
[0021] Step S26: Data supplement, using the generator to generate simulation data to fill in the original data.
[0022] Furthermore, in step S3, the construction of the legal knowledge assisted learning clustering model specifically includes the following steps:
[0023] Step S31: define the multivariate fusion distance, which is expressed as follows:
[0024] ;
[0025] Among them, i and j represent the index of the data vector, and represent the i-th data vector and the j-th data vector respectively, represents the multivariate fusion distance function, represents the multivariate fusion distance between the i-th data vector and the j-th data vector, Indicates the modulus length, Indicates taking the maximum value, represents the angle between the i-th data vector and the j-th data vector, represents the cosine function, represents the multivariate fusion weight, Indicates taking the Euclidean distance;
[0026] Step S32: Calculate the matching degree, which is expressed as follows:
[0027] ;
[0028] in, represents the matching function, represents the matching degree between the i-th data point and the j-th data point, represents the continuous product symbol, Represents the d-th data dimension value of the ith data point, Represents the d-th data dimension value of the j-th data point, and represents the matching coefficient, represents the average multivariate fusion distance between all data vectors;
[0029] Step S33: Design a cluster center update mechanism, which is expressed as follows:
[0030] ;
[0031] Among them, k represents the index of data clustering, represents the updated position of the kth cluster center, It means taking the data vector with the closest distance. represents the position of the kth cluster center before updating, represents the total number of data vectors of the kth cluster, represents the index of the data vector of the kth cluster, Indicates The location of the data vector, represents the average position of all data vectors before the k-th cluster update;
[0032] Step S34: Generate legal knowledge-assisted learning clusters, calculate the matching degree between each data vector and each cluster center, assign the data vector to the cluster with the highest matching degree, update the cluster center, recalculate the matching degree, assign the data vector again, and repeat the data vector assignment and cluster center update until the cluster no longer changes;
[0033] Step S35: Setting the abnormality detection distance, as shown below:
[0034] ;
[0035] in, represents the anomaly detection distance, represents the multivariate fusion distance between the two data vectors with the greatest distance, Represents the multivariate fusion distance between the two closest data vectors;
[0036] Step S36: Clear abnormal vectors. When the number of data vectors in a cluster is less than 30% of the average number of data vectors in all clusters, and any data vector of this cluster does not exist in other clusters within the abnormal detection distance range, this cluster is set as an abnormal cluster and the data vectors in this cluster are cleared.
[0037] Furthermore, in step S4, the legal knowledge assisted learning collects user data, uses the legal knowledge assisted learning clustering model, performs clustering operations, and obtains all data in the same cluster as the user; pushes legal knowledge resources corresponding to legal knowledge points in the cluster whose learning progress is less than 50% to the user; and issues an abnormal warning when the user belongs to an abnormal cluster.
[0038] The legal knowledge assisted learning system based on big data provided by the present invention includes a data acquisition module, a data processing module, a legal knowledge assisted learning clustering model building module and a legal knowledge assisted learning module;
[0039] The data collection module collects the student's grade, learning behavior data, legal ability data, legal knowledge learning situation data and legal knowledge resources, and sends the data to the data processing module;
[0040] The data processing module receives the data sent by the data acquisition module, processes the data, and sends the data to the module for building a clustering model for legal knowledge auxiliary learning;
[0041] The module for constructing a clustering model for assisting legal knowledge learning receives data sent by the data processing module, constructs a clustering model for assisting legal knowledge learning, and sends the data to the module for assisting legal knowledge learning;
[0042] The legal knowledge auxiliary learning module receives data sent by the legal knowledge auxiliary learning clustering model construction module, collects user data, performs clustering, and obtains all data in the same cluster as the user; pushes legal knowledge resources corresponding to legal knowledge points in the cluster whose learning progress is less than 50% to the user; and issues an abnormal warning when the user belongs to an abnormal cluster.
[0043] The beneficial effects achieved by the present invention using the above scheme are as follows:
[0044] (1) In view of the problems of low data processing efficiency, low data quality and difficulty in effectively generating high-quality learning data in traditional legal knowledge-assisted learning methods, this scheme improves the quality of generated data by constructing a generator and a discriminator and designing a loss function, thereby increasing the diversity and representativeness of the data and alleviating the problem of data scarcity.
[0045] (2) In view of the problems of poor clustering effect, difficulty in accurately reflecting the complex relationship between data vectors, inflexible cluster center update mechanism and inappropriate abnormal data processing in traditional legal knowledge-assisted learning clustering methods, this scheme improves clustering by defining multivariate fusion distance, calculating matching degree, designing cluster center update mechanism, setting abnormal detection distance and removing abnormal vectors. This improves the clustering algorithm's ability to identify complex relationships between data vectors, can adapt to changes in data vectors more quickly, avoids the interference of abnormal data on clustering results, improves clustering accuracy and stability, and enhances the adaptability of clustering algorithms. BRIEF DESCRIPTION OF THE DRAWINGS
[0046] Figure 1 A schematic diagram of the legal knowledge assisted learning method based on big data provided by the present invention;
[0047] Figure 2 A schematic diagram of the legal knowledge auxiliary learning system based on big data provided by the present invention;
[0048] Figure 3 A schematic diagram of data processing;
[0049] Figure 4 A schematic diagram for constructing a clustering model for assisting learning of legal knowledge;
[0050] Figure 5 It is a schematic diagram of the effect of data processing;
[0051] Figure 6 Schematic diagram of the implementation effect of constructing a legal knowledge assisted learning clustering model.
[0052] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention. DETAILED DESCRIPTION
[0053] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.
[0054] In the description of the present invention, it should be understood that terms such as “upper”, “lower”, “front”, “back”, “left”, “right”, “top”, “bottom”, “inside” and “outside” indicating directions or positional relationships are based on the directions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore should not be understood as limiting the present invention.
[0055] Example 1, see Figure 1 The present invention provides a legal knowledge auxiliary learning method based on big data, which comprises the following steps:
[0056] Step S1: Data collection, collecting students' grade, learning behavior data, legal ability data, legal knowledge learning data and legal knowledge resources; the learning behavior data includes learning time, learning frequency and number of discussions; the legal ability data includes homework completion and test scores; the legal knowledge learning data refers to the students' learning progress of legal knowledge points; the legal knowledge resources are textbooks, courses, papers, cases, legal provisions and legal classics corresponding to each legal knowledge point;
[0057] Step S2: data processing, processing data by constructing data vectors, constructing generators, constructing discriminators, designing generator loss functions, designing discriminator loss functions and data supplementation;
[0058] Step S3: constructing a legal knowledge-assisted learning clustering model, by defining a multivariate fusion distance, calculating a matching degree, designing a cluster center update mechanism, generating legal knowledge-assisted learning clusters, setting anomaly detection distances, and removing anomaly vectors;
[0059] Step S4: Legal knowledge assists learning, clustering is performed, and all data in the same cluster as the user is obtained; legal knowledge resources corresponding to legal knowledge points in the cluster whose learning progress is less than 50% are pushed to the user; when the user belongs to an abnormal cluster, an abnormal warning is issued.
[0060] Example 2, see Figure 1 and Figure 3 This embodiment is based on the above embodiment, and the data processing specifically includes the following steps:
[0061] Step S21: constructing a data vector, constructing the collected data into a data vector in the format of: the student's grade, study time, study frequency, number of discussions, homework completion, test scores, and learning progress of legal knowledge points;
[0062] Step S22: construct a generator, the generator includes an input layer, a hidden layer and an output layer; the input layer receives a noise vector; the hidden layer performs an upsampling operation; the output layer uses a hyperbolic tangent function to map the output;
[0063] Step S23: construct a discriminator, which includes a convolution layer, a fully connected layer and an output layer; the convolution layer performs a one-dimensional convolution operation and then maps the output by a linear rectification function; the fully connected layer maps the extracted features to the output space to reduce the dimension of the input data; the output layer uses a sigmoid function to output a value between 0 and 1, indicating the probability that the input data is real data;
[0064] Step S24: Design the generator loss function, which is expressed as follows:
[0065] ;
[0066] in, represents the loss value of the generator, represents the simulation data generated by the generator, represents a generator, Represents the generator based on noisy data and the conditional vector The generated simulation data represents the distribution of simulation data generated by the generator, represents taking the logarithm, represents the discriminator, It means taking the expected value. and represents the generation loss weight, l represents the layer index of the discriminator, Indicates that the discriminator uses simulated data at layer l The extracted features, It means that the discriminator is trained on the real data at layer l The extracted features, It means taking the square of L2 norm, represents the distribution of real data, Indicates that the discriminator is inputting real data And the conditional vector The identification result at Indicates that the discriminator is inputting simulated data And the conditional vector The identification result at that time;
[0067] Step S25: Design the discriminator loss function, expressed as follows:
[0068] ;
[0069] in, represents the loss value of the discriminator, represents the identification loss weight, d represents the dimension index of the data vector, and D represents the maximum dimension of the data vector. Represents the value of the real data in the dth dimension, Represents the value of the simulation data in the dth dimension;
[0070] Step S26: Data supplement, using the generator to generate simulation data to fill in the original data.
[0071] By performing the above operations, in order to address the problems of low data processing efficiency, low data quality and difficulty in effectively generating high-quality learning data in traditional legal knowledge assisted learning methods, this solution improves the quality of generated data by constructing a generator and a discriminator and designing a loss function, thereby improving the diversity and representativeness of the data and alleviating the problem of data scarcity.
[0072] Example 3, see Figure 1 and Figure 4 This embodiment is based on the above embodiment, and the construction of the legal knowledge assisted learning clustering model specifically includes the following steps:
[0073] Step S31: define the multivariate fusion distance, which is expressed as follows:
[0074] ;
[0075] Among them, i and j represent the index of the data vector, and represent the i-th data vector and the j-th data vector respectively, represents the multivariate fusion distance function, represents the multivariate fusion distance between the i-th data vector and the j-th data vector, Indicates the modulus length, Indicates taking the maximum value, represents the angle between the i-th data vector and the j-th data vector, represents the cosine function, represents the multivariate fusion weight, Indicates taking the Euclidean distance;
[0076] Step S32: Calculate the matching degree, which is expressed as follows:
[0077] ;
[0078] in, represents the matching function, represents the matching degree between the i-th data point and the j-th data point, represents the continuous product symbol, Represents the d-th data dimension value of the ith data point, Represents the d-th data dimension value of the j-th data point, and represents the matching coefficient, represents the average multivariate fusion distance between all data vectors;
[0079] Step S33: Design a cluster center update mechanism, which is expressed as follows:
[0080] ;
[0081] Among them, k represents the index of data clustering, represents the updated position of the kth cluster center, It means taking the data vector with the closest distance. represents the position of the kth cluster center before updating, represents the total number of data vectors of the kth cluster, represents the index of the data vector of the kth cluster, Indicates The location of the data vector, represents the average position of all data vectors before the k-th cluster update;
[0082] Step S34: Generate legal knowledge-assisted learning clusters, calculate the matching degree between each data vector and each cluster center, assign the data vector to the cluster with the highest matching degree, update the cluster center, recalculate the matching degree, assign the data vector again, and repeat the data vector assignment and cluster center update until the cluster no longer changes;
[0083] Step S35: Setting the abnormality detection distance, as shown below:
[0084] ;
[0085] in, represents the anomaly detection distance, represents the multivariate fusion distance between the two data vectors with the greatest distance, Represents the multivariate fusion distance between the two closest data vectors;
[0086] Step S36: Clear abnormal vectors. When the number of data vectors in a cluster is less than 30% of the average number of data vectors in all clusters, and any data vector of this cluster does not exist in other clusters within the abnormal detection distance range, this cluster is set as an abnormal cluster and the data vectors in this cluster are cleared.
[0087] By performing the above operations, this solution improves clustering by defining multivariate fusion distance, calculating matching degree, designing cluster center update mechanism, setting abnormal detection distance and clearing abnormal vectors to solve the problems of poor clustering effect, difficulty in accurately reflecting complex relationships between data vectors, inflexible cluster center update mechanism and inappropriate abnormal data processing in traditional legal knowledge assisted learning clustering methods. This improves the clustering algorithm's ability to identify complex relationships between data vectors, can adapt to changes in data vectors more quickly, avoids interference of abnormal data on clustering results, improves clustering accuracy and stability, and enhances the adaptability of clustering algorithms.
[0088] Example 4, see Figure 1 This embodiment is based on the above embodiment. The legal knowledge assisted learning collects the user's grade, study time, study frequency, number of discussions, homework completion, test scores and learning progress of legal knowledge points, and uses the legal knowledge assisted learning clustering model to perform clustering operations to obtain all data in the same cluster as the user; pushes legal knowledge resources corresponding to legal knowledge points in the cluster whose learning progress is less than 50% to the user; and issues an abnormal warning when the user belongs to an abnormal cluster.
[0089] Example 5, see Figure 1 and Figure 2 , this embodiment is based on the above embodiment, the legal knowledge assisted learning system based on big data provided by the present invention includes a data acquisition module, a data processing module, a legal knowledge assisted learning clustering model building module and a legal knowledge assisted learning module;
[0090] The data collection module collects the student's grade, learning behavior data, legal ability data, legal knowledge learning situation data and legal knowledge resources, and sends the data to the data processing module;
[0091] The data processing module receives the data sent by the data acquisition module, processes the data, and sends the data to the module for building a clustering model for legal knowledge auxiliary learning;
[0092] The module for constructing a clustering model for assisting legal knowledge learning receives data sent by the data processing module, constructs a clustering model for assisting legal knowledge learning, and sends the data to the module for assisting legal knowledge learning;
[0093] The legal knowledge auxiliary learning module receives data sent by the legal knowledge auxiliary learning clustering model construction module, collects user data, performs clustering, and obtains all data in the same cluster as the user; pushes legal knowledge resources corresponding to legal knowledge points in the cluster whose learning progress is less than 50% to the user; and issues an abnormal warning when the user belongs to an abnormal cluster.
[0094] Example 6, see Figure 5 and Figure 6 This embodiment is based on the above embodiment, and the specific implementation effects of the data processing and construction of the legal knowledge assisted learning clustering model are as follows:
[0095] Data processing, training the generator and discriminator, and obtaining the loss values of the training process of the generator and discriminator. The training loss of the model gradually decreases with the increase of the number of training times, and the overall performance of the model is improved;
[0096] A legal knowledge-assisted learning clustering model was constructed, and data vectors were clustered to obtain normal clusters and abnormal clusters. The number of data vectors in normal clusters was greater than that in abnormal clusters. Through clustering processing, each data vector was matched with a legal knowledge-assisted learning cluster.
[0097] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.
[0098] While the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that many changes, modifications, substitutions and variations can be made to the embodiments without departing from the principles and spirit of the invention.
[0099] The present invention and its embodiments are described above, which is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if ordinary technicians in the field are inspired by it and design structural methods and embodiments similar to the technical solution without creativity without departing from the purpose of the invention, they should all fall within the protection scope of the present invention.
Claims
1. A legal knowledge assisted learning method based on big data, characterized in that: The method comprises the following steps: Step S1: Data collection, collecting students’ grade, learning behavior data, legal ability data, legal knowledge learning data and legal knowledge resources; Step S2: data processing, processing data by constructing data vectors, constructing generators, constructing discriminators, designing generator loss functions, designing discriminator loss functions and data supplementation; Step S3: constructing a legal knowledge-assisted learning clustering model, by defining a multivariate fusion distance, calculating a matching degree, designing a cluster center update mechanism, generating legal knowledge-assisted learning clusters, setting anomaly detection distances, and removing anomaly vectors; Among them, the multivariate fusion distance is defined as follows: ； Among them, i and j represent the index of the data vector, and represent the i-th data vector and the j-th data vector respectively, represents the multivariate fusion distance function, represents the multivariate fusion distance between the i-th data vector and the j-th data vector, Indicates the modulus length, Indicates taking the maximum value, represents the angle between the i-th data vector and the j-th data vector, represents the cosine function, represents the multivariate fusion weight, Indicates taking the Euclidean distance; Step S4: Legal knowledge assists learning, clustering is performed, and all data in the same cluster as the user is obtained; legal knowledge resources corresponding to legal knowledge points in the cluster whose learning progress is less than 50% are pushed to the user; when the user belongs to an abnormal cluster, an abnormal warning is issued.
2. The big data-based legal knowledge assisted learning method according to claim 1 is characterized by: In step S3, the construction of the legal knowledge assisted learning clustering model specifically includes the following steps: Step S31: define multivariate fusion distance; Step S32: Calculate the matching degree, which is expressed as follows: ； in, represents the matching function, represents the matching degree between the i-th data point and the j-th data point, d represents the dimension index of the data vector, and D represents the maximum dimension of the data vector. represents the continuous product symbol, Represents the d-th data dimension value of the ith data point, Represents the d-th data dimension value of the j-th data point, and represents the matching coefficient, represents the average multivariate fusion distance between all data vectors; Step S33: Design a cluster center update mechanism, which is expressed as follows: ； Among them, k represents the index of data clustering, represents the updated position of the kth cluster center, It means taking the data vector with the closest distance. represents the position of the kth cluster center before updating, represents the total number of data vectors of the kth cluster, represents the index of the data vector of the kth cluster, Indicates The location of the data vector, represents the average position of all data vectors before the k-th cluster update; Step S34: Generate legal knowledge-assisted learning clusters, calculate the matching degree between each data vector and each cluster center, assign the data vector to the cluster with the highest matching degree, update the cluster center, recalculate the matching degree, assign the data vector again, and repeat the data vector assignment and cluster center update until the cluster no longer changes; Step S35: Setting the abnormality detection distance, as shown below: ； in, represents the anomaly detection distance, represents the multivariate fusion distance between the two data vectors with the greatest distance, Represents the multivariate fusion distance between the two closest data vectors; Step S36: Clear abnormal vectors. When the number of data vectors in a cluster is less than 30% of the average number of data vectors in all clusters, and any data vector of this cluster does not exist in other clusters within the abnormal detection distance range, this cluster is set as an abnormal cluster and the data vectors in this cluster are cleared.
3. The big data-based legal knowledge assisted learning method according to claim 1 is characterized by: In step S2, the data processing specifically includes the following steps: Step S21: constructing a data vector, constructing the collected data into a data vector in the format of: the student's grade, study time, study frequency, number of discussions, homework completion, test scores, and learning progress of legal knowledge points; Step S22: construct a generator, the generator includes an input layer, a hidden layer and an output layer; the input layer receives a noise vector; the hidden layer performs an upsampling operation; the output layer uses a hyperbolic tangent function to map the output; Step S23: construct a discriminator, which includes a convolution layer, a fully connected layer and an output layer; the convolution layer performs a one-dimensional convolution operation and then maps the output by a linear rectification function; the fully connected layer maps the extracted features to the output space to reduce the dimension of the input data; the output layer uses a sigmoid function to output a value between 0 and 1, indicating the probability that the input data is real data; Step S24: Design the generator loss function, which is expressed as follows: ； in, represents the loss value of the generator, represents the simulation data generated by the generator, represents a generator, Represents the generator based on noisy data and the conditional vector The generated simulation data represents the distribution of simulation data generated by the generator, represents taking the logarithm, represents the discriminator, It means taking the expected value. and represents the generation loss weight, l represents the layer index of the discriminator, Indicates that the discriminator uses simulated data at layer l The extracted features, Indicates that the discriminator is on the real data at layer l The extracted features, It means taking the square of L2 norm, represents the distribution of real data, Indicates that the discriminator is inputting real data and the conditional vector The identification result at Indicates that the discriminator is inputting simulated data and the conditional vector The identification result at that time; Step S25: Design the discriminator loss function, expressed as follows: ； in, represents the loss value of the discriminator, represents the identification loss weight, Represents the value of the real data in the dth dimension, Represents the value of the simulation data in the dth dimension; Step S26: Data supplement, using the generator to generate simulation data to fill in the original data.
4. The big data-based legal knowledge assisted learning method according to claim 1 is characterized by: In step S1, the data collection is to collect the student's grade, learning behavior data, legal ability data, legal knowledge learning situation data and legal knowledge resources; the learning behavior data includes learning time, learning frequency and number of discussions participated; the legal ability data includes homework completion and test scores; the legal knowledge learning situation data refers to the student's learning progress of legal knowledge points; the legal knowledge resources are textbooks, courses, papers, cases, legal provisions and legal classics corresponding to each legal knowledge point; In step S4, the legal knowledge assisted learning collects the user's grade, study time, study frequency, number of discussions participated in, homework completion, test scores and learning progress of legal knowledge points, and uses the legal knowledge assisted learning clustering model to perform clustering operations to obtain all data in the same cluster as the user; pushes legal knowledge resources corresponding to legal knowledge points in the cluster whose learning progress is less than 50% to the user; and issues an abnormal warning when the user belongs to an abnormal cluster.
5. A legal knowledge assisted learning system based on big data, used to implement the legal knowledge assisted learning method based on big data as described in any one of claims 1 to 4, characterized in that: It includes data collection module, data processing module, legal knowledge assisted learning clustering model building module and legal knowledge assisted learning module; The data collection module collects the student's grade, learning behavior data, legal ability data, legal knowledge learning situation data and legal knowledge resources, and sends the data to the data processing module; The data processing module receives the data sent by the data acquisition module, processes the data, and sends the data to the module for building a clustering model for legal knowledge assisted learning; The module for constructing a clustering model for assisting legal knowledge learning receives data sent by the data processing module, constructs a clustering model for assisting legal knowledge learning, and sends the data to the module for assisting legal knowledge learning; The legal knowledge auxiliary learning module receives data sent by the legal knowledge auxiliary learning clustering model construction module, collects user data, performs clustering, and obtains all data in the same cluster as the user; pushes legal knowledge resources corresponding to legal knowledge points in the cluster whose learning progress is less than 50% to the user; and issues an abnormal warning when the user belongs to an abnormal cluster.